akaaafk commited on
Commit
0e3b555
·
verified ·
1 Parent(s): f8c975d

Add files using upload-large-folder tool

Browse files
Files changed (20) hide show
  1. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_component_scaffold_v2_16x_4k.json +0 -0
  2. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_component_scaffold_v2_16x_4k.md +336 -0
  3. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_derivation_scaffold_16x_4k.json +0 -0
  4. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_derivation_scaffold_16x_4k.md +17 -0
  5. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_concise_hints_groot_4k.json +0 -0
  6. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_concise_hints_groot_4k.md +39 -0
  7. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_val_component_scaffold_v2_16x_4k.json +0 -0
  8. benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_val_component_scaffold_v2_16x_4k.md +316 -0
  9. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/summary.json +15 -0
  10. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/train.jsonl +0 -0
  11. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/val.jsonl +0 -0
  12. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/summary.json +13 -0
  13. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/train.jsonl +0 -0
  14. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/val.jsonl +0 -0
  15. benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v9_atomic_with_prior/summary.json +13 -0
  16. benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v2/summary.json +8 -0
  17. benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v2/train.jsonl +0 -0
  18. benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v3_clean_prompt/summary.json +7 -0
  19. benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v3_clean_prompt/train.jsonl +0 -0
  20. benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v4_expanded_clean/summary.json +8 -0
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_component_scaffold_v2_16x_4k.json ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_component_scaffold_v2_16x_4k.md ADDED
@@ -0,0 +1,336 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # IOLing Hint Diversity Probe
2
+
3
+ - generated: `2026-07-10T23:35:13.066706+00:00`
4
+ - model: `Qwen/Qwen3-4B-Instruct-2507`
5
+ - samples: `32`
6
+ - positive samples: `32`
7
+ - exact samples: `31`
8
+
9
+ ## Method Summary
10
+
11
+ | split | method | records | samples | records + | records exact | sample + | sample exact | mean best | format | unique boxes |
12
+ | --- | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
13
+ | val | derivation_scaffold_hint | 2 | 32 | 2 | 2 | 1.000 | 0.969 | 1.000 | 1.000 | 3 |
14
+
15
+ ## Positive Samples
16
+
17
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `0`
18
+
19
+ - score: `1.0`
20
+ - matched: `b.1, b.2`
21
+ - missed: ``
22
+
23
+ ```text
24
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
25
+ ```
26
+
27
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `1`
28
+
29
+ - score: `1.0`
30
+ - matched: `b.1, b.2`
31
+ - missed: ``
32
+
33
+ ```text
34
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
35
+ ```
36
+
37
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `2`
38
+
39
+ - score: `1.0`
40
+ - matched: `b.1, b.2`
41
+ - missed: ``
42
+
43
+ ```text
44
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
45
+ ```
46
+
47
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `3`
48
+
49
+ - score: `1.0`
50
+ - matched: `b.1, b.2`
51
+ - missed: ``
52
+
53
+ ```text
54
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
55
+ ```
56
+
57
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `4`
58
+
59
+ - score: `1.0`
60
+ - matched: `b.1, b.2`
61
+ - missed: ``
62
+
63
+ ```text
64
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
65
+ ```
66
+
67
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `5`
68
+
69
+ - score: `1.0`
70
+ - matched: `b.1, b.2`
71
+ - missed: ``
72
+
73
+ ```text
74
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
75
+ ```
76
+
77
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `6`
78
+
79
+ - score: `1.0`
80
+ - matched: `b.1, b.2`
81
+ - missed: ``
82
+
83
+ ```text
84
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
85
+ ```
86
+
87
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `7`
88
+
89
+ - score: `1.0`
90
+ - matched: `b.1, b.2`
91
+ - missed: ``
92
+
93
+ ```text
94
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
95
+ ```
96
+
97
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `8`
98
+
99
+ - score: `1.0`
100
+ - matched: `b.1, b.2`
101
+ - missed: ``
102
+
103
+ ```text
104
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
105
+ ```
106
+
107
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `9`
108
+
109
+ - score: `1.0`
110
+ - matched: `b.1, b.2`
111
+ - missed: ``
112
+
113
+ ```text
114
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
115
+ ```
116
+
117
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `10`
118
+
119
+ - score: `1.0`
120
+ - matched: `b.1, b.2`
121
+ - missed: ``
122
+
123
+ ```text
124
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
125
+ ```
126
+
127
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `11`
128
+
129
+ - score: `1.0`
130
+ - matched: `b.1, b.2`
131
+ - missed: ``
132
+
133
+ ```text
134
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
135
+ ```
136
+
137
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `12`
138
+
139
+ - score: `1.0`
140
+ - matched: `b.1, b.2`
141
+ - missed: ``
142
+
143
+ ```text
144
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
145
+ ```
146
+
147
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `14`
148
+
149
+ - score: `1.0`
150
+ - matched: `b.1, b.2`
151
+ - missed: ``
152
+
153
+ ```text
154
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
155
+ ```
156
+
157
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `15`
158
+
159
+ - score: `1.0`
160
+ - matched: `b.1, b.2`
161
+ - missed: ``
162
+
163
+ ```text
164
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
165
+ ```
166
+
167
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `0`
168
+
169
+ - score: `1.0`
170
+ - matched: `c.1, c.2`
171
+ - missed: ``
172
+
173
+ ```text
174
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
175
+ ```
176
+
177
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `1`
178
+
179
+ - score: `1.0`
180
+ - matched: `c.1, c.2`
181
+ - missed: ``
182
+
183
+ ```text
184
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
185
+ ```
186
+
187
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `2`
188
+
189
+ - score: `1.0`
190
+ - matched: `c.1, c.2`
191
+ - missed: ``
192
+
193
+ ```text
194
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
195
+ ```
196
+
197
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `3`
198
+
199
+ - score: `1.0`
200
+ - matched: `c.1, c.2`
201
+ - missed: ``
202
+
203
+ ```text
204
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
205
+ ```
206
+
207
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `4`
208
+
209
+ - score: `1.0`
210
+ - matched: `c.1, c.2`
211
+ - missed: ``
212
+
213
+ ```text
214
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
215
+ ```
216
+
217
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `5`
218
+
219
+ - score: `1.0`
220
+ - matched: `c.1, c.2`
221
+ - missed: ``
222
+
223
+ ```text
224
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
225
+ ```
226
+
227
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `6`
228
+
229
+ - score: `1.0`
230
+ - matched: `c.1, c.2`
231
+ - missed: ``
232
+
233
+ ```text
234
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
235
+ ```
236
+
237
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `7`
238
+
239
+ - score: `1.0`
240
+ - matched: `c.1, c.2`
241
+ - missed: ``
242
+
243
+ ```text
244
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
245
+ ```
246
+
247
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `8`
248
+
249
+ - score: `1.0`
250
+ - matched: `c.1, c.2`
251
+ - missed: ``
252
+
253
+ ```text
254
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
255
+ ```
256
+
257
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `9`
258
+
259
+ - score: `1.0`
260
+ - matched: `c.1, c.2`
261
+ - missed: ``
262
+
263
+ ```text
264
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
265
+ ```
266
+
267
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `10`
268
+
269
+ - score: `1.0`
270
+ - matched: `c.1, c.2`
271
+ - missed: ``
272
+
273
+ ```text
274
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
275
+ ```
276
+
277
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `11`
278
+
279
+ - score: `1.0`
280
+ - matched: `c.1, c.2`
281
+ - missed: ``
282
+
283
+ ```text
284
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
285
+ ```
286
+
287
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `12`
288
+
289
+ - score: `1.0`
290
+ - matched: `c.1, c.2`
291
+ - missed: ``
292
+
293
+ ```text
294
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
295
+ ```
296
+
297
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `13`
298
+
299
+ - score: `1.0`
300
+ - matched: `c.1, c.2`
301
+ - missed: ``
302
+
303
+ ```text
304
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
305
+ ```
306
+
307
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `14`
308
+
309
+ - score: `1.0`
310
+ - matched: `c.1, c.2`
311
+ - missed: ``
312
+
313
+ ```text
314
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
315
+ ```
316
+
317
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `15`
318
+
319
+ - score: `1.0`
320
+ - matched: `c.1, c.2`
321
+ - missed: ``
322
+
323
+ ```text
324
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
325
+ ```
326
+
327
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `13`
328
+
329
+ - score: `0.5`
330
+ - matched: `b.2`
331
+ - missed: `b.1`
332
+
333
+ ```text
334
+ b.1: om-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
335
+ ```
336
+
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_derivation_scaffold_16x_4k.json ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v6_val_derivation_scaffold_16x_4k.md ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # IOLing Hint Diversity Probe
2
+
3
+ - generated: `2026-07-10T23:31:43.624011+00:00`
4
+ - model: `Qwen/Qwen3-4B-Instruct-2507`
5
+ - samples: `32`
6
+ - positive samples: `0`
7
+ - exact samples: `0`
8
+
9
+ ## Method Summary
10
+
11
+ | split | method | records | samples | records + | records exact | sample + | sample exact | mean best | format | unique boxes |
12
+ | --- | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
13
+ | val | derivation_scaffold_hint | 2 | 32 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 32 |
14
+
15
+ ## Positive Samples
16
+
17
+ No positive samples found.
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_concise_hints_groot_4k.json ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_concise_hints_groot_4k.md ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # IOLing Hint Diversity Probe
2
+
3
+ - generated: `2026-07-10T23:38:59.581555+00:00`
4
+ - model: `Qwen/Qwen3-4B-Instruct-2507`
5
+ - samples: `345`
6
+ - positive samples: `1`
7
+ - exact samples: `0`
8
+
9
+ ## Method Summary
10
+
11
+ | split | method | records | samples | records + | records exact | sample + | sample exact | mean best | format | unique boxes |
12
+ | --- | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
13
+ | train | anti_nearmiss | 9 | 36 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 28 |
14
+ | train | groot_1x4 | 9 | 36 | 1 | 0 | 0.028 | 0.000 | 0.037 | 1.000 | 36 |
15
+ | train | groot_anchor_iid | 9 | 9 | 0 | 0 | 0.000 | 0.000 | 0.000 | 0.889 | 9 |
16
+ | train | iid | 9 | 36 | 0 | 0 | 0.000 | 0.000 | 0.000 | 0.917 | 27 |
17
+ | train | redacted_solution_hint | 9 | 18 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 18 |
18
+ | train | solver_checklist | 9 | 36 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 30 |
19
+ | train | unit_audit | 9 | 36 | 0 | 0 | 0.000 | 0.000 | 0.000 | 0.917 | 28 |
20
+ | val | anti_nearmiss | 6 | 24 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 23 |
21
+ | val | groot_1x4 | 6 | 24 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 23 |
22
+ | val | groot_anchor_iid | 6 | 6 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 6 |
23
+ | val | iid | 6 | 24 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 24 |
24
+ | val | redacted_solution_hint | 6 | 12 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 12 |
25
+ | val | solver_checklist | 6 | 24 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 22 |
26
+ | val | unit_audit | 6 | 24 | 0 | 0 | 0.000 | 0.000 | 0.000 | 1.000 | 24 |
27
+
28
+ ## Positive Samples
29
+
30
+ ### `train` `iol-2016-individual-p5-sub-a` `groot_1x4` sample `1`
31
+
32
+ - score: `0.3333333333333333`
33
+ - matched: `a.1`
34
+ - missed: `a.2, a.3`
35
+
36
+ ```text
37
+ a.1: is it in my house?; a.2: in the pot; a.3: in the field
38
+ ```
39
+
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_val_component_scaffold_v2_16x_4k.json ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/reports/ioling_hint_diversity/Qwen_Qwen3-4B-Instruct-2507__qwen3_4b_manual_v7_val_component_scaffold_v2_16x_4k.md ADDED
@@ -0,0 +1,316 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # IOLing Hint Diversity Probe
2
+
3
+ - generated: `2026-07-10T23:39:41.687115+00:00`
4
+ - model: `Qwen/Qwen3-4B-Instruct-2507`
5
+ - samples: `32`
6
+ - positive samples: `30`
7
+ - exact samples: `29`
8
+
9
+ ## Method Summary
10
+
11
+ | split | method | records | samples | records + | records exact | sample + | sample exact | mean best | format | unique boxes |
12
+ | --- | --- | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: | ---: |
13
+ | val | derivation_scaffold_hint | 2 | 32 | 2 | 2 | 0.938 | 0.906 | 1.000 | 0.938 | 4 |
14
+
15
+ ## Positive Samples
16
+
17
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `0`
18
+
19
+ - score: `1.0`
20
+ - matched: `b.1, b.2`
21
+ - missed: ``
22
+
23
+ ```text
24
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
25
+ ```
26
+
27
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `1`
28
+
29
+ - score: `1.0`
30
+ - matched: `b.1, b.2`
31
+ - missed: ``
32
+
33
+ ```text
34
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
35
+ ```
36
+
37
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `2`
38
+
39
+ - score: `1.0`
40
+ - matched: `b.1, b.2`
41
+ - missed: ``
42
+
43
+ ```text
44
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
45
+ ```
46
+
47
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `3`
48
+
49
+ - score: `1.0`
50
+ - matched: `b.1, b.2`
51
+ - missed: ``
52
+
53
+ ```text
54
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
55
+ ```
56
+
57
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `4`
58
+
59
+ - score: `1.0`
60
+ - matched: `b.1, b.2`
61
+ - missed: ``
62
+
63
+ ```text
64
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
65
+ ```
66
+
67
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `5`
68
+
69
+ - score: `1.0`
70
+ - matched: `b.1, b.2`
71
+ - missed: ``
72
+
73
+ ```text
74
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
75
+ ```
76
+
77
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `6`
78
+
79
+ - score: `1.0`
80
+ - matched: `b.1, b.2`
81
+ - missed: ``
82
+
83
+ ```text
84
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
85
+ ```
86
+
87
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `7`
88
+
89
+ - score: `1.0`
90
+ - matched: `b.1, b.2`
91
+ - missed: ``
92
+
93
+ ```text
94
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
95
+ ```
96
+
97
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `8`
98
+
99
+ - score: `1.0`
100
+ - matched: `b.1, b.2`
101
+ - missed: ``
102
+
103
+ ```text
104
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
105
+ ```
106
+
107
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `9`
108
+
109
+ - score: `1.0`
110
+ - matched: `b.1, b.2`
111
+ - missed: ``
112
+
113
+ ```text
114
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
115
+ ```
116
+
117
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `10`
118
+
119
+ - score: `1.0`
120
+ - matched: `b.1, b.2`
121
+ - missed: ``
122
+
123
+ ```text
124
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
125
+ ```
126
+
127
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `11`
128
+
129
+ - score: `1.0`
130
+ - matched: `b.1, b.2`
131
+ - missed: ``
132
+
133
+ ```text
134
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
135
+ ```
136
+
137
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `12`
138
+
139
+ - score: `1.0`
140
+ - matched: `b.1, b.2`
141
+ - missed: ``
142
+
143
+ ```text
144
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
145
+ ```
146
+
147
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `14`
148
+
149
+ - score: `1.0`
150
+ - matched: `b.1, b.2`
151
+ - missed: ``
152
+
153
+ ```text
154
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
155
+ ```
156
+
157
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `15`
158
+
159
+ - score: `1.0`
160
+ - matched: `b.1, b.2`
161
+ - missed: ``
162
+
163
+ ```text
164
+ b.1: öm-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
165
+ ```
166
+
167
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `0`
168
+
169
+ - score: `1.0`
170
+ - matched: `c.1, c.2`
171
+ - missed: ``
172
+
173
+ ```text
174
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
175
+ ```
176
+
177
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `1`
178
+
179
+ - score: `1.0`
180
+ - matched: `c.1, c.2`
181
+ - missed: ``
182
+
183
+ ```text
184
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
185
+ ```
186
+
187
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `2`
188
+
189
+ - score: `1.0`
190
+ - matched: `c.1, c.2`
191
+ - missed: ``
192
+
193
+ ```text
194
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
195
+ ```
196
+
197
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `3`
198
+
199
+ - score: `1.0`
200
+ - matched: `c.1, c.2`
201
+ - missed: ``
202
+
203
+ ```text
204
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
205
+ ```
206
+
207
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `4`
208
+
209
+ - score: `1.0`
210
+ - matched: `c.1, c.2`
211
+ - missed: ``
212
+
213
+ ```text
214
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
215
+ ```
216
+
217
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `5`
218
+
219
+ - score: `1.0`
220
+ - matched: `c.1, c.2`
221
+ - missed: ``
222
+
223
+ ```text
224
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
225
+ ```
226
+
227
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `6`
228
+
229
+ - score: `1.0`
230
+ - matched: `c.1, c.2`
231
+ - missed: ``
232
+
233
+ ```text
234
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
235
+ ```
236
+
237
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `8`
238
+
239
+ - score: `1.0`
240
+ - matched: `c.1, c.2`
241
+ - missed: ``
242
+
243
+ ```text
244
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
245
+ ```
246
+
247
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `9`
248
+
249
+ - score: `1.0`
250
+ - matched: `c.1, c.2`
251
+ - missed: ``
252
+
253
+ ```text
254
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
255
+ ```
256
+
257
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `10`
258
+
259
+ - score: `1.0`
260
+ - matched: `c.1, c.2`
261
+ - missed: ``
262
+
263
+ ```text
264
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
265
+ ```
266
+
267
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `11`
268
+
269
+ - score: `1.0`
270
+ - matched: `c.1, c.2`
271
+ - missed: ``
272
+
273
+ ```text
274
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
275
+ ```
276
+
277
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `12`
278
+
279
+ - score: `1.0`
280
+ - matched: `c.1, c.2`
281
+ - missed: ``
282
+
283
+ ```text
284
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
285
+ ```
286
+
287
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `13`
288
+
289
+ - score: `1.0`
290
+ - matched: `c.1, c.2`
291
+ - missed: ``
292
+
293
+ ```text
294
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
295
+ ```
296
+
297
+ ### `val` `iol-2015-individual-p1-sub-c` `derivation_scaffold_hint` sample `15`
298
+
299
+ - score: `1.0`
300
+ - matched: `c.1, c.2`
301
+ - missed: ``
302
+
303
+ ```text
304
+ c.1: fete nimbo ngámbi; c.2: yànparo tarumba yenówe fete asàr tàxwo tambaroy
305
+ ```
306
+
307
+ ### `val` `iol-2015-individual-p1-sub-b` `derivation_scaffold_hint` sample `13`
308
+
309
+ - score: `0.5`
310
+ - matched: `b.2`
311
+ - missed: `b.1`
312
+
313
+ ```text
314
+ b.1: om-pöhualli-om-öme; b.2: cen-tzontli-on-näuh-pöhualli-om-mahtlactli-on-nähui
315
+ ```
316
+
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/summary.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_rl_manual_expanded_clean_v14",
3
+ "parent_data": "data/rl/ioling_qwen3_4b_manual_v13_expanded_clean",
4
+ "manual_expansion_ledger": "data/manual_overrides/train_expansion_v14.json",
5
+ "train_count": 123,
6
+ "val_count": 24,
7
+ "total_count": 147,
8
+ "expansion_count": 35,
9
+ "train_source_problem_count": 15,
10
+ "val_source_problem_count": 3,
11
+ "source_overlap": [],
12
+ "prompt_style": "strict_keyed_atomic_clean_v14",
13
+ "scoring_overrides": "data/manual_overrides/scoring_overrides_v1.json",
14
+ "note": "Adds 35 manually checked Kimbundu, Hakhun, and Terena tasks; validation is unchanged."
15
+ }
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v14_expanded_clean/val.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/summary.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_rl_atomic_v8",
3
+ "parent_data": "data/rl/ioling_qwen3_4b_manual_v7_concise",
4
+ "train_count": 38,
5
+ "val_count": 24,
6
+ "total_count": 62,
7
+ "train_source_problem_count": 6,
8
+ "val_source_problem_count": 3,
9
+ "source_overlap": [],
10
+ "prompt_style": "strict_keyed_atomic_v8",
11
+ "include_prior_answers": false,
12
+ "note": "Every row targets one manually PDF-checked answer unit; source-problem split is preserved."
13
+ }
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v8_atomic/val.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/rl/ioling_qwen3_4b_manual_v9_atomic_with_prior/summary.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_rl_atomic_with_prior_v9",
3
+ "parent_data": "data/rl/ioling_qwen3_4b_manual_v7_concise",
4
+ "train_count": 38,
5
+ "val_count": 24,
6
+ "total_count": 62,
7
+ "train_source_problem_count": 6,
8
+ "val_source_problem_count": 3,
9
+ "source_overlap": [],
10
+ "prompt_style": "strict_keyed_atomic_with_prior_v9",
11
+ "include_prior_answers": true,
12
+ "note": "Every row targets one manually PDF-checked answer unit; later items receive explicitly paired earlier targets and verified answers; source-problem split is preserved."
13
+ }
benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v2/summary.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_selected_student_traces_v2",
3
+ "rows": 61,
4
+ "source_problems": 9,
5
+ "manual_expansion_rows": 23,
6
+ "all_production_exact": true,
7
+ "manual_review": "Every completion checked against target-specific facts; expansion transcribed directly from rendered official pages."
8
+ }
benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v2/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v3_clean_prompt/summary.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_selected_student_traces_v3_clean_prompt",
3
+ "parent_data": "data/sft/ioling_selected_student_traces_v2",
4
+ "rows": 61,
5
+ "source_problems": 9,
6
+ "all_production_exact": true
7
+ }
benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v3_clean_prompt/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
benchmark/IOL/ioling_hf/sft/ioling_selected_student_traces_v4_expanded_clean/summary.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "ioling_selected_student_traces_v4_expanded_clean",
3
+ "parent_data": "data/sft/ioling_selected_student_traces_v3_clean_prompt",
4
+ "rows": 80,
5
+ "source_problems": 11,
6
+ "manual_expansion_rows": 19,
7
+ "all_production_exact": true
8
+ }